Upskilling vs reskilling: the difference and how to do it

Both terms describe targeted training, but with different goals. With upskilling, the role stays; it changes. An accountant learns to work with automated reconciliations and data analysis. A service technician learns to diagnose equipment remotely. With reskilling, someone changes role: an employee from manual data entry becomes the contact person for customer data, a production worker becomes a quality inspector.

Automation, artificial intelligence and new regulations change tasks faster than people change professions. Many tasks do not vanish completely but change shape. That increases the need for upskilling in almost every role. At the same time, some roles lose their core. For these people, reskilling is the alternative to separation — and for the company often the cheaper and faster route to skilled staff who already know the organisation, its customers and its processes.

One example of a binding skills requirement is AI literacy: since 2 February 2025 the EU AI Act has required providers and deployers of AI systems to take measures so that their staff have a sufficient level of AI literacy (Art. 4).

Both start by looking ahead: Which tasks will exist in two to three years, which will change, which will disappear? From this you can derive future roles and their skills. Comparing them with today's situation — for example using a skills matrix — shows who can get there through upskilling and for whom a change of role makes sense.

Honesty is crucial here. Reskilling programmes launched without real target positions create hope and then disappointment. Better: name concrete positions, disclose the requirements and make selection transparent.

Upskilling can usually be built into existing learning paths: short units on new tools, practice tasks in daily work, a certificate at the end. Reskilling needs more structure: its own learning path over several months, learning time within working hours, a buddy in the target department, practical phases and clearly defined milestones.

Change creates uncertainty. Anyone who learns that their role is changing fundamentally first asks: What does this mean for me? Clear communication, a real perspective and confidence that learning is welcome rather than seen as weakness matter more than any course format. Managers play the lead role — they must protect learning time and recognise progress.

Certificates help the individual too: a verifiable certificate for a newly acquired skill keeps its value, even beyond the company.

Upskilling in procurement: a company introduces a new tendering system and wants to analyse supplier data more closely. Buyers stay in their role but, over a six-week learning path, learn to use the system, the basics of data analysis and how to handle AI-supported suggestions. Each step ends with a task on real cases; at the end, the manager confirms the new skill level.

Reskilling in administration: automation removes a large part of manual document entry. Instead of cutting jobs, some employees are qualified to handle exceptions and look after suppliers. The learning path lasts several months and includes subject knowledge, conversation skills and a practical phase in the target department with an experienced buddy. The positions in the target role are fixed in advance; selection is based on interest and aptitude.

What both examples share: the goal is concrete, learning time is scheduled, and the result is visibly confirmed.